A high fidelity simulation of the North American Eastern Interconnect known as ERGIS–Eastern Renewable Generation Integration Study–indicates that the system could continue to function in the year 2026, even if as much as 30% of its annual electricity generation and consumption was produced using variable power sources like the wind and the sun.
At the high end of the penetration levels studied, there are times when the wind and sun together provide as much as 52% of the total demand and as little as 10%. Sunrise and sunset routinely become periods of intense system response. On some days, 140 GW of production must shift from solar and wind to gas or coal fired generators during a period lasting less than six hours.
A press conference was held in late August to announce the release of a report that describes the creation and initial results from ERGIS.
During the press conference, Aaron Bloom, the ERGIS study project manager, stated that the project results summary describes four specific scenarios chosen with inputs from a Technical Review Committee. He also acknowledged that there were many more scenarios and constraining assumptions that would be of interest to decision makers. NREL expects and encourages requests from other research groups to use the tools developed during the project to explore a wide range of possibilities and answer questions that the initial team had not considered.
Genesis Of ERGIS
The Eastern Interconnect is the power industry’s name for the world’s largest, most complex and economically critical system. It is the electrical power grid that serves the geographic half of the US from eastern sections of Montana and New Mexico–excluding most of Texas–to the Atlantic coast plus the central and eastern provinces of Canada with the exception of Quebec.
That service territory is one of the world’s most densely populated and productive regions; virtually every activity in the region depends on the reliable flow of electricity from thousands of generating sources to hundreds of millions of customers. The system has been planned and constructed over the past century to handle an ever changing load and to be as resilient as economically feasible in the event of unplanned events like generator failures, transmission line faults and severe weather.
During the past decade there has been increasing momentum to provide electricity generated by capturing natural energy flows from the wind and sun. These generators reduce overall system fuel consumption and produce no CO2 when they are operating, but their output is almost completely dependent on variable weather conditions and the predictable rotation of the Earth.
The driving forces for added sources of weather dependent power and the need to study their operational impact were: efforts to reduce CO2 emissions from electricity production, “state renewable portfolio standards (RPS), federal policies affecting the tax structures of wind and PV projects, renewable technology advancements, and cost decreases” pg. 1
Starting in 2012, the National Renewable Energy Laboratory was tasked with building a high fidelity model of the system that could help decision makers understand the operational impacts for the system of a rapidly increasing share of electricity from weather-dependent sources.
The NREL team had access to the high performance computing capabilities of NREL’s Energy Systems Integration Facility, which includes a system called Peregrine. That massive array of processors and storage is considered to be one of the 50 fastest computers in the world and is the highest ranking one that is dedicated to studies of renewable energy and energy efficiency integration.
Even though the team was using one of the most powerful computer systems in the world, they quickly realized they needed to develop a number of innovative problem solving techniques to enable solution of complex, multi-input equations in 5-minute intervals for an entire year without requiring months of runtime per scenario. In addition to the computing optimization work, the team made a number of simplifying assumptions that have the potential for increasing the gap between simulation and reality.
Examples Of Important Assumptions
At the high end, the southeast US is assumed to be able to obtain 15% of its wind electricity from wind farms in the Southwest Power Pool while Florida is assumed to be able to install 1.5 kW/person of distributed solar capacity – 50% more than any other area in the study and 10% of the installed capacity in the state. The Virginia-Carolinas area was assumed to be able to obtain 80% of its wind from offshore installations.
Note: Exactly one offshore wind installation has been completed in the US. It is expected to be operational by the end of 2016. It has a total generating capacity of 30 MWe and it is off of the coast of Block Island, RI.
The study used summarized weather data from 2006 as an input for both electricity demand and production capability from the wind and sun. Researchers acknowledged that using a single year of weather data imposed a fidelity constraint by possibly overlooking severe weather events like lengthy heat waves or the Polar Vortex.
Coal plant retirement assumptions were based on announced plans, old age and low utilization rates. All nuclear plants operating in 2010 were assumed to continue operating with the exception of Crystal River, Kewaunee, Vermont Yankee and Oyster Creek. When not in a planned outage, remaining nuclear plants were assumed to be operated at their nameplate capacity and not be used to follow any loads.
All coal, nuclear and natural gas plant retirements were assumed to be replaced with new natural gas fired generation.
Coal and natural gas fired power plants were assumed to be adequately supplied by fuel at prices predicted by the Energy Information Agency 2014 Annual Energy Outlook for the year 2026. Pipeline constraints were not modeled.
The entire Eastern Interconnect was assumed to be controlled by a single mathematical model for commitments and dispatch. The reality is that there are numerous regions within the EI that have a limited ability–both physically and procedurally–to transfer power into other regions. The single formula assumption often resulted in the model deciding to move power over much longer distances than is usually the case.
Value And Limitations Of High Fidelity Modeling
Unlike the facile statements about a 100% renewable energy future by renewable energy advocates like the Solutions Project, the ERGIS report and the associated animations should help responsible policy makers understand that it is not easy to maintain grid reliability as variable power sources like the wind and the sun play a growing role. The system becomes increasingly dynamic and uses up some of its installed resilience to handle such expected events as sunset.
The study team was honest enough to provide various forms of the following warning to policy makers in all three of the study’s main communications products–the final report, the executive summary and the press release announcing the availability of the study.
However, we did not investigate whether transmission and generation operators will have sufficient incentives to provide the necessary ramping, energy, and capacity services for futures like the ones we studied. While ERGIS shows it is technically possible to balance periods of instantaneous VG [variable generation] penetrations that exceed 50% for the EI, the ability of the real system to realize these futures may depend more on regulatory policy, market design, and operating procedures.
They did not note, however, that there weren’t any members of the technical review committee (TRC) who could help the researchers quantify the scale of “sufficient incentives” that might be required. The 20 member TRC included two professional renewable energy industry lobbyists–one from the American Wind Energy Association and one from the Solar Energy Industry Association–neither of whom has a documented technical or operational background.
People who will end up footing the bills and potentially suffering the consequences if the reduced resilience results in power limitations or outages should demand to know more about the required “sufficient incentives.” The entities who are assumed to provide the needed services will expect to be paid, sometimes quite handsomely. When read carefully, the ERGIS shows that increasing renewable penetration, even if the cost of the power generators themselves falls a bit, isn’t going to be easy, cheap or without risk to reliability.